C.S.C. SEKHAR

1
DETERMINANTS OF STATE FRAGILITY A
CROSS-COUNTRY ANALYSIS

• C.S.C. SEKHAR
• Institute of Economic Growth
• University of Delhi Enclave
• Delhi - 110 007
• e-mail sekharcsc_at_yahoo.com csekhar_at_iegindia.org
  

2
Heterogeneity of the Fragile States
3
Motivation

• Is there a discernible pattern in state fragility
  / collapse across countries?
• Are there clear, tell-tale symptoms of state
  collapse?
• What are the factors that avert state collapse or
  make reconstruction possible in some states but
  hinder the same in other states?

4
Objectives

• To identify the factors and interlinkages among
  sub-systems of a state that have a bearing on
  state fragility

5
Important Literature State Failure

• Dependency Theory
• Economic and political crises due to capitalist
  nature of the economies and exploitation by
  developed countries
• Prescription Delink from the global system and
  adopt state-controlled strategy
• Neo-liberal Theories - Crises due to overly
  interventionist state policies, misallocation of
  resources and subsidization of inefficiency
• Patrimonialism/clientelism Granting and
  receiving favours based on individualistic /
  community norms

6
Analytical Framework

• Combination of Douglas North (1994) s
  institutional matrix approach and a Weberian
  Lockean notions of the state
• Institutional Matrix Approach
• Institutions are at the core of the evolution of
  a society
• Institutions determine the type of organizations
  that come into existence

7
Core functions of the state (as adopted in the
study -Milliken and Krause 2002, CSRC 2006,
Lister 2007)

• Providing security from external aggression
• Provision of public goods and services
• Promoting welfare through economic growth and
  equitable distribution
• Providing representation to peoples aspirations

8
Hypotheses of the study

• Simultaneity among the institutions / processes
  in economic, political and social domains
• Vulnerability in one sphere affects and is
  affected by vulnerability in other spheres

9
Simultaneity of Economic, Political and Social
Vulnerability

• PV SV
• EV

10
Econometric Model
11
Econometric Model (contd)

• EV Vulnerability arising from low level and
  high inequality of income
• PV Inability to provide security, public goods,
  ensure human rights etc. Lack of legitimacy
• SV Group grievance and human suffering due to
  lack of social cohesion

12
Econometric Model (contd)

• XEV, XPV, XSV are the other pre-determined/exogeno
  us variables affecting economic, political and
  social vulnerability respectively.
• Method of Estimation 3SLS to account for
  contemporaneous correlation among errors across
  equations
• Model Evaluation PRMSE, Theils U Statistic,
  Bias proportion of U and Variance proportion of U

13
Data Sources and Methods

• Four sources used FSI, WDI, WGI and CPI
• Failed States Index (FSI) 2007, Fund for Peace
• Each Country is rated for 12 indicators 4
  social, 2 economic and 6 political
• Rating done on a scale of 0-10. 0lowest
  intensity 10-highest intensity
• Total vulnerability of a country is the sum of
  all the scores
• Higher the aggregate score, greater is the
  vulnerability
• Variables drawn from the source EV, PV, SV

14
Data Sources and Methods (contd)

• World Development Indicators, World Bank
• Standard definitions of WDI
• Variables drawn External Debt, Per Capita
  Income, Health Infrastructure
• Worldwide Governance Indicators (WGI), World Bank
  Institute
• Country performance as percentile rank
• Higher the value, better is the governance
• Variables drawn Voice and Accountability,
  Stability and Lack of Violence, Rule of Law

15
Data Sources and Methods (contd)

• Corruption Perceptions Index (CPI), Transparency
  International
• Perceived corruption of a country as a score
• Country with lowest score is the one where
  corruption is perceived to be the highest
• Variables drawn Corruption
• 5) Dummy variables for Social Fragmentation,
  Policy Uncertainty, Capacity of the Institutions,
  Proximity to a Failed / Fragile State and War

16
Results
17
Model Evaluation

• Difference between the actual and estimated
  vulnerability (TV) was equal to or less than 10
  in only 11 out of the 149 countries studied
• Theils U, Bias and Variance proportions of U
  also very close to zero indicating good
  performance of the model

18
Conclusions

• Simultaneity among economic, political and social
  vulnerability
• Social vulnerability (SV) affects both EV and PV
  significantly
• PV has a significant effect on SV but not on EV
• Only a single dimension of EV affects either PV
  or SV significantly
• Results are also supported by a comparison of
  some of the fragile states experiences with
  those of stable states

19
Policy Implications

• Addressing social vulnerability and upholding
  social cohesion should receive priority in the
  most fragile states
• Political initiatives should follow to promote
  economic growth and reduce inequalities

20
Limitations of the Model

• Model is based on cross-country regressions at a
  single point in time, mainly because of data
  constraints.
• Results are subject to the limitations of the
  data on the indicators of state failure,
  governance etc. and need careful interpretation.

21
THANK YOU